US10103924B1 - Phase correction of channelizer output without multipliers - Google Patents

Abstract

A method and system for phase correction in a receiver receives a plurality of samples and filters the samples using a polyphase FIR filter. The samples are simultaneously shifted, centered and channelized using a MUX/negate process and a fast Fourier transform (FFT).

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is a continuation of U.S. application Ser. No. 14/692,040, filed Apr. 21, 2015, entitled, “Phase Correction of Channelizer Output without Multipliers,” which is incorporated herein by reference in its entirety for all purposes.

BACKGROUND 1. Technical Field

This disclosure is related to digital channelizers and, more particularly, to digital channelizers having phase correction.

2. Discussion of Related Art

Frequency Division Multiple Access (FDMA) provides access for multiple users to a single channel by subdividing the channel into multiple sub-channels. Each user is allocated one or a subset of all the channels, and the user's signal is used to modulate a carrier of the allocated sub-channel. A single-channel receiver may be used to demodulate a single assigned channel, or a multichannel receiver may demodulate a selected range of channels.

In an analog system, each receiver has a local oscillator (LO) tuned to the frequency of the channel to be demodulated. The LO signal is multiplied by the received signal to generate an intermediate frequency (IF) signal, which is applied to a band-pass filter (BPF). The BPF has a passband selected to filter out all signals except the channel to be received. The output signal from the BPF can be digitized and processed subsequently, if desired.

SUMMARY

According to one aspect, a method for phase correction in a receiver is provided. The method includes: receiving a plurality of samples; filtering the samples using a polyphase FIR filter; and simultaneously shifting, centering and channelizing the samples using a MUX/negate process and a fast Fourier transform (FFT).

In some exemplary embodiments, the samples are matched to the channels, and the MUX/negate process and FFT are applied in accordance with
y _k(t)=[FFT[P _Mt((n−tM)mod N)]*e ^2πjMtk/N ]*e ^−πjMtk/N;
where N=number of channels, M=decimation factor, k=channel index, t=decimated time, and P_Mt and n represent the polyphase FIR filter with the FFT being applied over n.

In some exemplary embodiments, the samples are shifted by one half of a channel relative to the channels, and the MUX/negate process and FFT are applied in accordance with
y _k(t)=[OFFT[P _Mt(n−tM)mod N)]*e ^{2πjMt(k+1/2)/N} ]*e ^{−2πjMt(k+1/2)/N};
where N=number of channels, M=decimation factor, k=channel index, t=decimated time, and P_Mt and n represent the polyphase FIR filter with the FFT being applied over n.

According to another aspect, a receiver system with phase correction is provided. The receiver system includes a receiver for receiving a plurality of samples. The receiver system also includes a polyphase filter for filtering the samples, A processor simultaneously shifts, centers and channelizes the samples using a MUX/negate process and a fast Fourier transform (FFT).

In some exemplary embodiments, the samples are matched to the channels, and the MUX/negate process and FFT are applied in accordance with
y _k(t)=[FFT[P _Mt((n−tM)mod N]*e ^2πjMtk/N ]*e ^−πjMtk/N;
where N=number of channels, M=decimation factor, k=channel index, t=decimated time, and P_Mt and n represent the polyphase filter with the FFT being applied over n.

In some exemplary embodiments, the samples are shifted by one half of a channel relative to the channels, and the MUX/negate process and FFT are applied in accordance with
y _k(t)=[OFFT[P _Mt((n−tM)mod N)]*e ^{2πjMt(k+1/2)/N} ]*e ^{−2πjMt(k+1/2)/N};
where N=number of channels, M=decimation factor, k=channel index, t=decimated time, and P_Mt and n represent the polyphase filter with the FFT being applied over n.

BRIEF DESCRIPTION OF THE DRAWINGS

The present disclosure is further described in the detailed description which follows, in reference to the noted plurality of drawings by way of non-limiting examples of embodiments of the present disclosure, in which like reference numerals represent similar parts throughout the several views of the drawings.

FIG. 1 includes a schematic block diagram of a conventional receiver.

FIG. 2 includes a schematic block diagram of a conventional polyphase channelizer.

FIGS. 3A and 3B include schematic system block diagrams which illustrate conventional channelizer processing and channelizer processing according to exemplary embodiments, respectively, using the example of real input 4/3 oversampling and a 14-channel channelizer.

FIGS. 4A and 4B include frequency spectra for real-input 4/3 oversampled 14-channel channelizer without phase compensation and with phase compensation, respectively, according to exemplary embodiments.

FIGS. 5A and 5B include schematic block diagrams of a channelizer implementation having frequency complex multiplies and phasor lookup tables and a channelizer implementation without phasor multiplies, respectively, according to some exemplary embodiments.

DETAILED DESCRIPTION

FIG. 1 includes a schematic block diagram of a conventional receiver. Referring to FIG. 1, conventional receiver 100 includes a local oscillator (LO) 105 having frequency ω. A signal mixer 100 mixes the amplified incoming FDMA signal 106 with the LO 105 output to produce a signal 107 with a fixed intermediated frequency (IF). An analog BPF 111 selects the desired channel and an analog-to-digital converter (ADC) 115 converts the resulting output to digital form by sampling the analog signal at an appropriated frequency. Generally, the sampling frequency is at least twice the channel bandwidth to satisfy the Nyquist requirement.

The sampled digital data, x(n), is bandshifted digitally by digital detector 116 by multiplying with a phasor e^{−j2π(kn/N)} denoted by W_Nkn, where k denotes the channel selected by the receiver. When there are multiple channels contained in the received signal, then one such receiver is needed for each of the channels. The resulting signal is low-pass filtered using a digital low pass filter 117 and then downsampled. All of the processes following digitization of the band-passed IF signal can be performed with a digital computer or specialized circuits well known in the art.

FIG. 2 includes a schematic block diagram of a conventional polyphase channelizer. Referring to FIG. 2, in a conventional polyphase approach to digital channelization, M channels of bandwidth B are received simultaneously in a FDMA signal 106. After M channels, each of bandwidth B, are passed through an analog BPF having a passband of at least M×B (not shown), they are output to an A/D converter 140, which samples at some rate that is at least equal to the Nyquist rate (2 MB) for a signal of bandwidth MB. In FIG. 2, the data is sampled at 2 NB, where N is greater then or equal to M. It should be noted that N does not have to be greater than or equal to M. There are undersampled applications as well, and the present disclosure is applicable to those cases. The digital output x(n) is applied to a 2N pole commutator 141, which distributes the input data x(n) to 2N filters 150. Each filter 150 is updated once every 2N points. Filters 150 perform the channel extraction function. The time series output of filters 150 is applied to respective inputs of an FFT processor 160, which processes the data once every 2N points to produce 2N complex outputs of which M outputs are chosen, each representing the band-shifted subchannel signal at B Hz, the update rate of FFT processor 160. Only M outputs of FFT processor 160 are required since the sample rate 2 NB can be higher than the minimum required sample rate 2 MB. The filter updates at an M rate, but all FFT outputs are valid. For the example shown, only one-half of the outputs of the FFT are necessary due to Nyquist.

Typically, channelizers have integer m/n ratios for oversample such that the number of channels is divisible by the downsample rate. This is desirable so that the resulting signals are just a linear frequency rotation of the desired bandwidth. If the number of channels is not divisible by m, then the resulting signal may have frequency offsets which are non-linear, and has to be compensated for on a sample-by-sample basis. Without the compensation, the channelizer output is cannot be used for processing. Specifically, for oversampled channelizers, the phase offsets vary non-linearly, sample by sample, and channel by channel. According to the present disclosure, the non-linear offsets are compensated for, such that the resulting signal is frequency centered and usable for post processing. The use of channelizers and oversampled or undersampled channelizers is common where the number of channels is evenly divisible by the oversample rate. However, the present disclosure is directed to providing for processing with non-divisible downsample rates. This allows for more flexibility in the system of the disclosure for frequency planning more options for bandwidths.

Conventional frequency planning requires modification of radio frequency (RF) local oscillators (LO's) and analog-to-digital converter (ADC) sample rates to provide bandwidths appropriate for current digital channelizers. Achieving output channel bandwidths digitally requires additional signal processing. Thus, the exemplary embodiments set forth in the present disclosure provide downsampled channelized output channels with non-integer oversampling ratios relative to the number of channels. The approach of the disclosure provides digital channelization of a spectrum to non-integer ratios relative to the number of channels. The approach of the disclosure combines an interpolation filter with the digital channelizer and corrects for non-linear phase offsets on the output channels.

Frequency planning for radar systems requires the selection of RF mixers so that the downconverted spectrums can be sampled by ADCs to cover the required spectrum. With existing systems and/or hardware, additional flexibility in processing in ADC sample rates provides the ability to limit RF modification and/or complexity. For a sampling frequency that requires a non-integer oversampling rates relative to the number of channels, a straightforward digital signal processing approach results in a resource-intense implementation.

For example, a 4/3 oversample for a 16-channel channelizer requires the channelizer to receive 12 new samples (16×¾) per processing clock. A 4/3 oversample for a 14-channel channelizer requires the channelizer to receive 10.5 new samples per processing clock. FIGS. 3A and 3B include schematic system block diagrams which illustrate conventional channelizer processing and channelizer processing according to exemplary embodiments, respectively, using the example of real input 4/3 oversampling and a 14-channel channelizer. Referring to FIG. 3A, in the conventional configuration, a signal Fs is input to a 2×interpolation filter 200 at input 206. A 2× interpolated signal 2*Fs from interpolation filter 200 is input at 208 to a 4/3 oversampled 28-channel channelizer 202. Duplicate channels due to real input and interpolation are discarded and information for seven channels @ Fs*4/3*14 is output on 210 to a linear phase correction channel centering block 204. Phase-corrected and channel-centered information is output at 204.

Referring to FIG. 3B, the optimized digital signal processing implementation of the present disclosure is illustrated. Referring to FIG. 3B, a signal Fs is input at 224 to a 4/3 oversampled 14-channel channelizer 220. According to the exemplary embodiments, interpolation is combined and performed simultaneously at 220 with channelization. The resulting output at 226, which includes information for seven channels @ Fs*4/3*14, is applied to non-linear phase correction and channel centering block 222. Non-linear phase-corrected and channel-centered information is output at 228.

Thus, according to the present disclosure, non-integer oversampling or undersampling rates relative to the number of channels result in non-linear phase offsets in each of the channels. According to the exemplary embodiments, phase compensation is introduced to compensate for the non-linear phase offsets. In contrast, in conventional channelizers, the frequency spectrum would be rotated, i.e., linear phase offsets. FIGS. 4A and 4B include frequency spectra for real-input 4/3 oversampled 14-channel channelizer without phase compensation and with phase compensation, respectively, according to exemplary embodiments.

Thus, the exemplary embodiments provide digital channelization of a spectrum to non-integer ratios relative to the number of channels. This approach combines an interpolation filter with the digital channelizer and corrects for non-linear phase offsets on the output channels.

The following describes in more detail the channelizer 220 and non-linear phase correction channel centering block 222 of FIG. 3B, according to the exemplary embodiments, in greater detail.

The interpolating channelizer 220 implements a resampling filter for the filter portion of the channelizer 220. A resampling filter up-samples by L and decimates by M. Up-sampling is needed if L/M is not an integer. For example, if it is desirable to downsample the signal by 1.5, it is necessary to discard every 1.5 samples. This is accomplished by up-sampling by 2, and decimating by 3. The processing produces a fractional delay of ½ samples where needed by interpolating the data.

For a typical channelizer, new samples are shifted into the polyphase filter structure at a constant rate, which results in the linear nature of the phase offsets needed. When M does not equal N, this has the effect of multiplying a spectrum by a linear changing phase offset, which rotates the spectrum. In contrast, for the interpolating channelizer 220 of the exemplary embodiments, the new samples are shifted into the polyphase structure at a non-linear rate which results in non-linear phase offsets. The spectrum information is modified such that it does not represent the spectrum without compensating for the non-linear phase offsets.

For example, a channelizer where N=4 and downsample rate=3 (4/3 oversampled, L=1, M=3), there are 4 filter branches in the polyphase filter, which receive 3 new samples every downsampled clock rate. A channelizer where N=6 and downsample rate=4.5 (4/3 oversampled, L=2, M=9), there are 6 filter branches in the resampling polyphase filter which receive 4 and 5 samples every other clock. As with a resampling filter, the branches alternate between two different sets of coefficients every other clock.

The derivation of a channelizer starts with a mixer, a low pass filter, and a decimation. The channelizer equation is derived by use of the Noble identity, polyphase filter structures, and a fast Fourier transform (FFT).

N=number of channels

M=decimation factor

k=channel index

t=decimated time

h_n(m)=h(n+Nm)

x_Mt(n)=x(Mt+n)

Total taps in the filter=T*N

$y_k\left(t\right)=\left[\sum _{n=0}^{\mathrm{TAPS}-1}\left(e^{-2\pi j\left(t-n\right)k/N}*x\left(t-n\right)\right)*h\left(n\right)\right]\left[M\downarrow \right]$ $y_k\left(t\right)=\left[\sum _{n=0}^{N-1}\sum _{m=0}^{T-1}{x}_{Mt}\left(n-\mathrm{mN}\right)*h_n\left(m\right)*e^{-2\pi \mathrm{jnk}/N}\right]*e^{-2\pi j\mathrm{Mtk}/N}$

The first portion of the channelizer 220 matches the form of a standard polyphase filter. For a downsampled polyphase filter design, the polyphase branches are chosen such that the number of branches equals the downsample rate. This produces the most efficient structure as branches are independent of each other. For an oversampled or undersampled channelizer, the number of polyphase branches equals the number of channels in order to match the throughput requirement of the FFT portion of the channelizer. The result is the branches share data with other branches depending on the ratio of M and N. Also, due to the mismatched number of channels and the decimation rate, the FFT portion results in phase components which are variable in channels and variable linearly in time. Without applying the phase component, each of the output channels spectrums will be rotated and the centers of the channels are not at DC.

Similarly, an interpolated channelizer according to the exemplary embodiments is derived from a mixer and a resampling filter.

$y_k\left(t\right)=\left[\sum _{n=0}^{\mathrm{TAPS}-1}\left[\left(e^{-2\pi j\left(t-r\right)k/N}*x\left(t-n\right)\right)\left[L\uparrow \right]\right]*h\left(n\right)\right]\left[M\downarrow \right]$
It is convenient to form the summation in reverse order from TAP-1 to 0, so that the summation increments with t. At the end of the derivation, the branches of the polyphase map directly to the indexes of the FFT rather than in reverse order. This is equivalent to adding a time offset (t+TAPS-1). This is removed from the equation for convenience. The equation is modified by substituting −n for n, and replacing ℏ(n) with h (n)=h (TAPS-1-n).

$y_k\left(t\right)=\left[\sum _{n=0}^{\mathrm{TAPS}-1}\left[\left(e^{-2\pi j\left(t+n\right)k/N}*x\left(t+n\right)\right)\left[L\uparrow \right]\right]*\stackrel{\_}{h}\left(n\right)\right]\left[M\downarrow \right]$
Up-sampling by L results in:

$y_k\left(t\right)=[\sum _{n=0}^{\mathrm{TAPS}-1}{e}^{-2\pi j\left(\mathrm{floor}\left(t/L\right)+n\right)k/N}*x\left(\mathrm{floor}\left(t/L\right)+n\right)*\stackrel{\_}{h}\left(\mathrm{nL}+t\mathrm{mod}L\right)]\hspace{1em}\left[M\downarrow \right]$
Converting to polyphase filter of branches N allows the mixer to be constant for each polyphase branch.

$y_k\left(t\right)=\left[\sum _{n=0}^{N-1}\sum _{m=0}^{T-1}{e}^{-2\pi j\left(\mathrm{floor}\left(t/L\right)+\left(n+\mathrm{mN}\right)\right)k/N}*x\left(\mathrm{floor}\left(t/L\right)+\left(n+\mathrm{mN}\right)\right)*\stackrel{\_}{h}\left(\left(n+\mathrm{mN}\right)L=t\mathrm{mod}L\right)\right]\left[M\downarrow \right]$ $y_k\left(t\right)=\left[\sum _{n=0}^{N-1}\sum _{m=0}^{T-1}{e}^{-2\pi j\left(\mathrm{floor}\left(Mt/L\right)+\left(n+\mathrm{mN}\right)\right)k/N}*x\left(\mathrm{floor}\left(Mt/L\right)+\left(n+\mathrm{mN}\right)\right)*\stackrel{\_}{h}\left(\mathrm{nL}+\mathrm{mNL}+Mt\mathrm{mod}L\right)\right]$ $y_k\left(t\right)=\left[\sum _{n=0}^{N-1}\sum _{m=0}^{T-1}{e}^{-2\pi j\left(\mathrm{floor}\left(Mt/L\right)\right)k/N}*e^{-2\pi \mathrm{jnk}/N}*e^{-2\pi \mathrm{jmNk}/N}*x\left(\mathrm{floor}\left(Mt/L\right)+n+\mathrm{mN}\right)*\stackrel{\_}{h}\left(\mathrm{mNL}+\mathrm{nL}+Mt\mathrm{mod}L\right)\right]$ $y_k\left(t\right)=e^{-2\pi j\left(\mathrm{floor}\left(Mt/L\right)\right)k/N}*[\sum _{n=0}^{N-1}[\sum _{m=0}^{T-1}{x}_{\mathrm{floor}\left(Mt/L\right)t}\left(n+\mathrm{mN}\right)*{\stackrel{\_}{h}}_{\mathrm{nL}+Mt\mathrm{mod}L}\left(m\right)]*e^{-2\pi \mathrm{jnk}/N}]$

The result is a resampling filter whose branch results are the inputs into an FFT. The result is multiplied by a phase coefficient dependent on time and channel. The polyphase resampling filter shifts new data across each branch at a non-linear rate, represented by floor(Mt/L)t. This is the same rate non-linear amount that the phase coefficients change. The phase coefficients can also apply patent application 13-4463, and replace the non-linear phase multiply with a rotation of the branch outputs prior to the FFT. The derivation is similar for a half-channel shifted channelizer, to replace the non-linear phase multiply with a data rotation of the filter branch outputs.

Example: 4/3 oversampling for N=6

• • N=6; L=2; M=9

$y_k\left(t\right)=e^{-2\pi j\left(\mathrm{floor}\left(9t/2\right)\right)k/6}*[\sum _{n=0}^5[\sum _{m=0}^{T-1}{x}_{\mathrm{floor}\left(9t/2\right)t}\left(n+m6\right)*{\stackrel{\_}{h}}_{n2+9t\mathrm{mod}2}\left(m\right)]*e^{-2\pi \mathrm{jnk}/6}]$ $y_k\left(t\right)=e^{-2\pi j(\mathrm{floor}\left(4.5t\right)k/6}*[\sum _{n=0}^5[\sum _{m=0}^{T-1}{x}_{\mathrm{floor}\left(4.5t\right)t}\left(n+6m\right)*{\stackrel{\_}{h}}_{2n+t\mathrm{mod}2}\left(m\right)]*e^{-2\pi \mathrm{jnk}/6}]$

Up-converting x(t) by 2 yields x(0), 0, x(1), 0, x(2), 0, . . . . For L=2, there the filter is separated into 2*N=12 polyphase branches. The interpolation calculated is represented by ℏ_{2n+t mod 2}(m). For the samples for y_k(t+1), the nine samples are shifted into the filter (M=9), half of which are zeros. This results in the other half of the branches to be calculated. For every other clock, four then five samples of x(t) are shifted into the filter, and the calculated branches alternate. The results of the branches are the input into a six point FFT, which is then multiplied the non-linear coefficient.

Oversampling and/or half-channel-shifted channelizers output channels that are non-channel-centered and require a phasor multiply to center the channel. Phasor multiplies require a complex multiply and phasor lookup table per channel. For example, in the case of 16 channels, 16 multiples with 16 phasor RAMs are required. If multiplies are scarce, as is common with field-programmable gate array (FPGA) implementations, the output channels can remain uncorrected, and compensated for in the post processing. However, bands of interest are more easily interpreted when channel-centered, so correcting for the frequency rotation is desirable. According to the present disclosure, the phase compensation is applied to all channels simultaneously by re-mapping the samples after the filter of the channelizer. This replaces the different phase multiplies per channel, by a re-mapping (muxing) function, substituting muxes for complex multiplies, reducing utilization and power consumption. The resulting output is mathematically equivalent and channel-centered.

Thus, according to the exemplary embodiments, the resources required for channel centering of non-critically-sampled and/or half-channel-shifted digital channelizers are reduced. Frequency rotation is applied across all channels simultaneously with polyphase filter output reordering and/or negation, instead of a frequency complex multiply and phasor lookup table for each channel.

FIGS. 5A and 5B include schematic block diagrams of a channelizer implementation having frequency complex multiplies and phasor lookup tables and a channelizer implementation without phasor multiplies, respectively, according to some exemplary embodiments. Referring to FIG. 5A, system 300 receives input at switch 302, which selects polyphase finite-impulse-response (FIR) filters 304 one at a time to produce the channelized signal, which is then applied to FFT 306. The outputs of FFT 306 are channel-centered by complex multiplies 301 of phasors from phasor RAM or phasor look-up table 308. Referring to FIG. 5B, system 400 receives input at switch 402, which selects polyphase FIR filters 404 one at a time to produce the channelized signal, which is then applied to MUX/negate circuit 408 to center the channels. Please describe MUX/negate circuit 408 in detail here. The channel-centered signals from MUX/negate circuit 408 are then applied to FFT 406 for further processing. Hence, according to the exemplary embodiments illustrated in FIG. 5B, the complex multiplies 310 and phasor lookup table 308 of FIG. 5A are replaced by the simpler MUX/negate circuit 408 for centering the channels.

Thus, the present disclosure provides a resource-optimized approach for channel centering the output channels of a non-critically centered digital channelizer. The approach of the disclosure replaces complex multipliers and phasor lookup tables with re-order muxing and/or negation.

Additional details of the approach of the disclosure are described below.

Standard Channelizer Equation Detailed Description

For a polyphase channelizer, standard and ½ channel shifted:

• • T=Taps per branch
  • N=number of channels
  • M=decimation factor
  • k=channel index
  • t=decimated time
  • h_n(m)=h(n+Nm)
  • x_Mt(n)=x(Mt+n)
  • Total taps in the filter=T*N

Standard channelizer (t is constant within the summation, and h_n represents each branch of the polyphase filter):

$y_k\left(t\right)=\sum _{n=0}^{N-1}\sum _{m=0}^{T-1}x\left(Mt+n-\mathrm{mN}\right)*h_n\left(m\right)*e^{-2\pi j\left(n+Mt\right)k/N}$ $y_k\left(t\right)=\left[\sum _{n=0}^{N-1}\sum _{m=0}^{T-1}{x}_{Mt}\left(n-\mathrm{mN}\right)*h_n\left(m\right)*e^{-2\pi j\mathrm{nk}/N}\right]*e^{-2\pi \mathrm{jMtk}/N}$
Standard polyphase FIR filter where each branch of the polyphase can be represented by P_Mt(n)

$P_{Mt}\left(n\right)=\sum _{m=0}^{T-1}{x}_{Mt}\left(n-\mathrm{mN}\right)*h_n\left(m\right)$
Standard FFT equation

$\mathrm{FFT}\left[P\left(n\right)\right]=\sum _{n=0}^{N-1}P\left(n\right)*e^{-2\pi \mathrm{jnk}/N}$
Substituting for P_Mt(n)

$y_k\left(t\right)=\left[\sum _{n=0}^{N-1}{P}_{Mt}\left(n\right)*e^{-2\pi \mathrm{jnk}/N}\right]*e^{-2\pi \mathrm{jMtk}/N}$
Substituting FFT
y _k(t)=FFT[P _Mt(n)]*^{e−2πjMtk/N}
The phase offset required is different for each channel k, and varies over time every for samples t.
phase_k(t)=e ^−2πjMtk/N

Therefore, a channelizer is calculated by processing a polyphase filter by its branches, where each branch is the input into an FFT, where the output is multiplied by a time varying phase value. If M=N (critically sampled channelizer) then phase values simplify to 1.

The phase offset required can be can be canceled out by rotating the order of the output of the polyphase filter banks by t*M. The remapping can be represented by the new index as ((n−t*M) modulo N). The equation then results in a phase component. The phase component has the property where it is periodic over N, so ((n−t*M) modulo N) can be represented in the equation without the modulo as (n−t*M).

$\mathrm{FFT}\left[P\left(n\right)\right]=\sum _{n=0}^{N-1}P\left(n\right)*e^{-2\pi \mathrm{jnk}/N}=\sum _{n=0}^{N-1}P\left(\left(n-\mathrm{tM}\right)\mathrm{mod}N\right)*e^{-2\pi j\left(\left(n-\mathrm{tM}\right)\mathrm{mod}N\right)k/N}$ $\mathrm{FFT}\left[P\left(n\right)\right]=\sum _{n=0}^{N-1}P\left(\left(n-\mathrm{tM}\right)\mathrm{mod}N\right)*e^{-2\pi j\left(n-\mathrm{tM}\right)k/N}$ $\mathrm{FFT}\left[P\left(n\right)\right]=\sum _{n=0}^{N-1}P\left(\left(n-\mathrm{tM}\right)\mathrm{mod}N\right)*e^{-2\pi \mathrm{jnk}/N}*e^{2\pi \mathrm{jtMk}/N}$ $\mathrm{FFT}\left[P\left(n\right)\right]=e^{2\pi \mathrm{jMtk}/N}*\sum _{n=0}^{N-1}P\left(\left(n-\mathrm{tM}\right)\mathrm{mod}N\right)*e^{-2\pi \mathrm{jnk}/N}$ $\mathrm{FFT}\left[P\left(n\right)\right]=e^{2\pi \mathrm{jMtk}/N}*\mathrm{FFT}\left[P\left(\left(n-\mathrm{tM}\right)\mathrm{mod}N\right)\right]$
From the equation above, the equation of the standard channelizer is:
y _k(t)=FFT[P _Mt(n)]*e ^2πjMtk/N

Applying the rotation to the data results in:
y _k(t)=[FFT[P _Mt((n−tM)mod N)]*e ^{2πjMt(k+1/2)/N} ]*e ^−2πjMtk/N;
y _k(t)=FFT[P _Mt((n−tM)mod N)]

Half-Channel shifted channelizer:

$y_k\left(t\right)=\sum _{n=0}^{N-1}\sum _{m=0}^{T-1}{x}_{Mt}\left(m\right)*h_n\left(\mathrm{nN}-m\right)*e^{-2\pi j\left(n+Mt\right)\left(k+1/2\right)/N}$ $y_k\left(t\right)=\left[\sum _{n=0}^{N-1}\sum _{m=0}^{T-1}{x}_{Mt}\left(m\right)*h_n\left(\mathrm{nN}-m\right)*e^{-2\pi \mathrm{jn}\left(k+1/2\right)/N}\right]*e^{-2\pi \mathrm{jMt}\left(k+1/2\right)/N}$ $y_k\left(t\right)=\left[\sum _{n=0}^{N-1}\sum _{m=0}^{T-1}{x}_{Mt}\left(m\right)*h_n\left(\mathrm{nN}-m\right)*e^{-2\pi \mathrm{jn}\left(k+1/2\right)/N}\right]*e^{-2\pi \mathrm{jMt}\left(k+1/2\right)/N}$

Standard polyphase FIR filter where each branch of the polyphase can be represented by P(n)

$P_{Mt}\left(n\right)=\sum _{m=0}^{T-1}{x}_{Mt}\left(n-\mathrm{mN}\right)*h_n\left(m\right)$
Standard odd FFT equation:

$\mathrm{OFFT}\left[P\left(n\right)\right]=\sum _{n=0}^{N-1}P\left(n\right)*e^{-2\pi \mathrm{jn}\left(k+1/2\right)/N}$
Substituting for P(n):

$y_k\left(t\right)=\left[\sum _{n=0}^{N-1}{P}_{Mt}\left(n\right)*e^{-2\pi \mathrm{jn}\left(k+1/2\right)/N}\right]*e^{-2\pi \mathrm{jMt}\left(k+1/2\right)/N}$
Substituting odd FFT:
y _k(t)=OFFT[P _Mt(n)]*e ^{−2πjMt(k+1/2)/N}
The phase offset required is different for each channel k, and varies over time every for samples t.
phase_k(t)=e ^{−2πjMt(k+1/2)/N}

The phase offset required can be can be canceled out by rotating the order of the output of the polyphase filter banks by t*M. The remapping can be represented by the new index as ((n−t*M) modulo N). The phase component has the property where it is periodic over N, so ((n−t*M) modulo N) can be represented in the equation without the modulo as (n−t*M).

$\mathrm{OFFT}\left[P\left(n\right)\right]=\sum _{n=0}^{N-1}P\left(n\right)*e^{-2\pi \mathrm{jn}\left(k+1/2\right)/N}=\sum _{n=0}^{N-1}P\left(\left(n-\mathrm{tM}\right)\mathrm{mod}N\right)*e^{-2\pi j\left(\left(n-\mathrm{tM}\right)\mathrm{mod}N\right)\left(k+1/2\right)/N}$ $\mathrm{OFFT}\left[P\left(n\right)\right]=\sum _{n=0}^{N-1}P\left(\left(n-\mathrm{tM}\right)\mathrm{mod}N\right)*e^{-2\pi j\left(n-\mathrm{tM}\right)\left(k+1/2\right)/N}$ $\mathrm{OFFT}\left[P\left(n\right)\right]=\sum _{n=0}^{N-1}P\left(\left(n-\mathrm{tM}\right)\mathrm{mod}N\right)*e^{-2\pi j\left(n\right)\left(k+1/2\right)/N}*e^{-2\pi j\left(Mt\right)\left(k+1/2\right)/N}$ $\mathrm{OFFT}\left[P\left(n\right)\right]=e^{2\pi \mathrm{jMt}\left(k+1/2\right)/N}*\sum _{n=0}^{N-1}P\left(\left(n-\mathrm{tM}\right)\mathrm{mod}N\right)*e^{-2\pi \mathrm{jn}\left(k+1/2\right)/N}$ $\mathrm{OFFT}\left[P\left(n\right)\right]=e^{2\pi \mathrm{jMt}\left(k+1/2\right)/N}*\mathrm{OFFT}\left[P\left(\left(n-\mathrm{tM}\right)\mathrm{mod}N\right)\right]$

From the above, the equation of the standard odd channelizer is:
y _k(t)=OFFT[P _Mt(n)]*e ^{−2πjMt(k+1/2)/N}
Applying the rotation to the data results in:
y _k(t)=[OFFT[P _Mt((n−tM)mod N)]*e ^{2πjMt(k+1/2)/N} ]*e ^{−2πjMt(k+1/2)/N};
y _k(t)=OFFT[P _Mt((n−tM)mod N)]

The current method for compensating for linear offsets/frequency rotation is represented by the standard channelizer equations of each section, where the complex multiply is defined by e^−2πjMtk/N and e^{−2πjMt(k+1/2)/N} for even and odd channelizers respectively. The complex multiply is performed after the summation for each channel k.

A lookup table is generally used to provide the coefficients necessary. The size of the lookup table is determined by the ratio of M/N. In simple cases where M is a multiple of N, the coefficient reduces down to 1 and (−1)^t respectively.

Example

Standard channelizer:

$y_k\left(t\right)=\sum _{n=0}^{N-1}\sum _{m=0}^{T-1}{x}_{Mt}\left(n-\mathrm{mN}\right)*h_n\left(m\right)*e^{-2\pi j\left(n+Mt\right)k/N}$ $y_k\left(t\right)=\left[\sum _{n=0}^{N-1}\sum _{m=0}^{T-1}{x}_{Mt}\left(n-\mathrm{mN}\right)*h_n\left(m\right)*e^{-2\pi \mathrm{jnk}/N}\right]*e^{-2\pi \mathrm{jMtk}/N}$
The standard polyphase FIR filter where each branch of the polyphase can be represented by P(n) and standard FFT equation are shown below.

$P_{Mt}\left(n\right)=\sum _{m=0}^{T-1}{x}_{Mt}\left(n-\mathrm{mN}\right)*h_n\left(m\right)$ $\mathrm{FFT}\left[P\left(n\right)\right]=\sum _{n=0}^{N-1}P\left(n\right)*e^{-2\pi \mathrm{jnk}/N}$
After substituting P(n) and FFT:

$y_k\left(t\right)=\left[\sum _{n=0}^{N-1}{P}_{Mt}\left(n\right)*e^{-2\pi \mathrm{jnk}/N}\right]*e^{-2\pi \mathrm{jMtk}/N}$
For the example:

N=4

M=3

k=channel index

t=decimated time

h_n(m)=h(n+Nm)

The phase component is:
phase_k(t)=e ^−2πj3tk/4
The phase offset required is different for each channel k, and varies over time every for samples t.
The phase component is replaced by remapping of the data as described by ((n−t*M) modulo N).
Remapping is (n−t*3) modulo 4

	t = 0	t = 1	t = 2	t = 3
	Q(0) = P(0)	Q(0) = P(1)	Q(0) = P(2)	Q(0) = P(3)
	Q(1) = P(1)	Q(1) = P(2)	Q(1) = P(3)	Q(1) = P(0)
	Q(2) = P(2)	Q(2) = P(3)	Q(2) = P(0)	Q(2) = P(1)
	Q(3) = P(3)	Q(3) = P(0)	Q(3) = P(1)	Q(3) = P(2)

For half channel shifted example of the sample 4/3 oversampling, the muxing negation is shown.

t = 0	t = 1	t = 2	t = 3	t = 4	t = 5	t =6	t = 7
Q(0) = P(0)	Q(0) = −P(1)	Q(0) = P(2)	Q(0) = −P(3)	Q(0) = −P(0)	Q(0) = −P(1)	Q(0) = −P(2)	Q(0) = −P(3)
Q(1) = P(1)	Q(1) = −P(2)	Q(1) = P(3)	Q(1) = P(0)	Q(1) = −P(1)	Q(1) = P(2)	Q(1) = −P(3)	Q(1) = −P(0)
Q(2) = P(2)	Q(2) = −P(3)	Q(2) = −P(0)	Q(2) = P(1)	Q(2) = −P(2)	Q(2) = −P(3)	Q(2) = −P(0)	Q(2) = −P(1)
Q(3) = P(3)	Q(3) = P(0)	Q(3) = −P(1)	Q(3) = P(2)	Q(3) = −P(3)	Q(3) = −P(0)	Q(3) = P(1)	Q(3) = −P(2)

Various embodiments of the above-described systems and methods may be implemented in digital electronic circuitry, in computer hardware, firmware, and/or software. The implementation can be as a computer program product (i.e., a computer program tangibly embodied in an information carrier). The implementation can, for example, be in a machine-readable storage device and/or in a propagated signal, for execution by, or to control the operation of, data processing apparatus. The implementation can, for example, be a programmable processor, a computer, and/or multiple computers.

A computer program can be written in any form of programming language, including compiled and/or interpreted languages, and the computer program can be deployed in any form, including as a stand-alone program or as a subroutine, element, and/or other unit suitable for use in a computing environment. A computer program can be deployed to be executed on one computer or on multiple computers at one site.

Method steps can be performed by one or more programmable processors executing a computer program to perform functions of the invention by operating on input data and generating output. Method steps can also be performed by and an apparatus can be implemented as special purpose logic circuitry. The circuitry can, for example, be a FPGA (field programmable gate array) and/or an ASIC (application-specific integrated circuit). Modules, subroutines, and software agents can refer to portions of the computer program, the processor, the special circuitry, software, and/or hardware that implements that functionality.

Processors suitable for the execution of a computer program include, by way of example, both general and special purpose microprocessors, and any one or more processors of any kind of digital computer. Generally, a processor receives instructions and data from a read-only memory or a random access memory or both. The essential elements of a computer are a processor for executing instructions and one or more memory devices for storing instructions and data. Generally, a computer can include, can be operatively coupled to receive data from and/or transfer data to one or more mass storage devices for storing data (e.g., magnetic, magneto-optical disks, or optical disks).

Data transmission and instructions can also occur over a communications network. Information carriers suitable for embodying computer program instructions and data include all forms of non-volatile memory, including by way of example semiconductor memory devices. The information carriers can, for example, be EPROM, EEPROM, flash memory devices, magnetic disks, internal hard disks, removable disks, magneto-optical disks, CD-ROM, and/or DVD-ROM disks. The processor and the memory can be supplemented by, and/or incorporated in special purpose logic circuitry.

To provide for interaction with a user, the above described techniques can be implemented on a computer having a display device. The display device can, for example, be a cathode ray tube (CRT) and/or a liquid crystal display (LCD) monitor. The interaction with a user can, for example, be a display of information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer (e.g., interact with a user interface element). Other kinds of devices can be used to provide for interaction with a user. Other devices can, for example, be feedback provided to the user in any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback). Input from the user can, for example, be received in any form, including acoustic, speech, and/or tactile input.

The above described techniques can be implemented in a distributed computing system that includes a back-end component. The back-end component can, for example, be a data server, a middleware component, and/or an application server. The above described techniques can be implemented in a distributing computing system that includes a front-end component. The front-end component can, for example, be a client computer having a graphical user interface, a Web browser through which a user can interact with an example implementation, and/or other graphical user interfaces for a transmitting device. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), the Internet, wired networks, and/or wireless networks.

The system can include clients and servers. A client and a server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

Packet-based networks can include, for example, the Internet, a carrier internet protocol (IP) network (e.g., local area network (LAN), wide area network (WAN), campus area network (CAN), metropolitan area network (MAN), home area network (HAN)), a private IP network, an IP private branch exchange (IPBX), a wireless network (e.g., radio access network (RAN), 802.11 network, 802.16 network, general packet radio service (GPRS) network, HiperLAN), and/or other packet-based networks. Circuit-based networks can include, for example, the public switched telephone network (PSTN), a private branch exchange (PBX), a wireless network (e.g., RAN, Bluetooth, code-division multiple access (CDMA) network, time division multiple access (TDMA) network, global system for mobile communications (GSM) network), and/or other circuit-based networks.

The computing device can include, for example, a computer, a computer with a browser device, a telephone, an IP phone, a mobile device (e.g., cellular phone, personal digital assistant (PDA) device, laptop computer, electronic mail device), and/or other communication devices. The browser device includes, for example, a computer (e.g., desktop computer, laptop computer) with a World Wide Web browser (e.g., Microsoft® Internet Explorer® available from Microsoft Corporation, Mozilla® Firefox available from Mozilla Corporation). The mobile computing device includes, for example, a Blackberry®, iPAD®, iPhone® or other smartphone device.

Whereas many alterations and modifications of the disclosure will no doubt become apparent to a person of ordinary skill in the art after having read the foregoing description, it is to be understood that the particular embodiments shown and described by way of illustration are in no way intended to be considered limiting. Further, the subject matter has been described with reference to particular embodiments, but variations within the spirit and scope of the disclosure will occur to those skilled in the art. It is noted that the foregoing examples have been provided merely for the purpose of explanation and are in no way to be construed as limiting of the present disclosure.

Although the present disclosure has been described herein with reference to particular means, materials and embodiments, the present disclosure is not intended to be limited to the particulars disclosed herein; rather, the present disclosure extends to all functionally equivalent structures, methods and uses, such as are within the scope of the appended claims.

Claims (6)

The invention claimed is:

1. A method for digital channelization in a receiver, comprising:

receiving a plurality of samples associated with a plurality of channels;

shifting the samples into a polyphase finite impulse response (FIR) filter at a non-linear rate, generating non-linear phase offsets associated with each channel: and

simultaneously shifting, centering and channelizing the samples using a MUX/negate process and a fast Fourier transform (FFT) to correct for the non-linear phase offsets across all of the plurality of channels for each sample.

2. The method of claim 1, wherein the samples are matched to the channels, and the MUX/negate process and FFT are applied in accordance with

y _k(t)=[FFT[P _Mt((n−tM)mod N)]*e ^2πjMtk/N ]*e ^{−2πjMt(k+1/2)/N};

where N=number of channels, M=decimation factor, k=channel index, t=decimated time, and P_Mt and n represent the polyphase FIR filter with the FFT being applied over n.

3. The method of claim 1, wherein the samples are shifted by one half of a channel relative to the channels, and the MUX/negate process and FFT are applied in accordance with

y _k(t)=[OFFT[P _Mt((n−tM)mod N)]*e ^{2πjMt(k+1/2)/N} ]*e ^{−2πjMt(k+1/2)/N};

where N=number of channels, M=decimation factor, k=channel index, t=decimated time, and P_Mt and n represent the polyphase FIR filter with the FFT being applied over n.

4. A receiver system comprising:

a receiver for receiving a plurality of samples associated with a plurality of channels;

a polyphaser finite impulse response (FIR) filter the samples being shifted into the polyphase filter at a nonlinear rate and generating non-linear phase offsets associated with each channel: and

a processor for simultaneously shifting, centering and channelizing the samples using a MUX/negate process and a fast Fourier transform (FFT) to correct for the non-linear phase offsets across all of the plurality of channels for each sample.

5. The system of claim 4, wherein the samples are matched to the channels, and the MUX/negate process and FFT are applied in accordance with

y _k(t)=[FFT[P _Mt((n−tM)mod N)]*e ^2πjMtk/N ]*e ^−2πjMtk/N;

where N=number of channels, M=decimation factor, k=channel index, t=decimated time, and P_Mt and n represent the polyphase filter with the FFT being applied over n.

6. The system of claim 4, wherein the samples are shifted by one half of a channel relative to the channels, and the MUX/negate process and FFT are applied in accordance with

y _k(t)=[OFFT[P _Mt((n−tM)mod N)]*e ^{2πjMt(k+1/2)/N} ]*e ^{−2πjMt(k+1/2)/N};

where N=number of channels, M=decimation factor, k=channel index, t=decimated time, and P_Mt and n represent the polyphase filter with the FFT being applied over n.

Images